# hyperscaler capex — X 热门讨论 (2026-09-26 18:28 UTC)
## @0xTykoo (Tykoo) · 09-26 15:06 · ♥104 ↻19 💬0 红杉的一位合伙人 Pat Grady 昨天直接把他们和自己的一个 LP meeting（Boston College）的内容录制发了出来，信息密度非常高，强烈推荐大家去看一下：
- JEV 在过去的 7 天里从零增长到了 100 million 的收入。在此之前 The Information 刚刚爆出来，他们现在估值是在 10 billion。
- 现在 AI 公司的增速，比历史上的都要快很多。Instinct 发布以来，保持了 day over day 10% 的增长率。
- 200 到 700，指的是某一个 high-value knowledge work company 从去年年底 200M 增长到今年预测 700M 的收入。（后面那个 2 到 50 和 0 到 70%，他说还是跳过吧，可能还是有些敏感。）
- 在 AI 公司，现在有一个连融两轮的行业惯例或风气：把"陪你建公司的"和"只出钱的"拆开。他拿红杉内部的数据（过去 12 个月里的 7 个案例）来举例：
红杉作为陪建的那一方进场，平均的投后估值是 1.1 亿；而仅仅一个月后，只出钱的那一方给的下一轮平均估值就达到了 34 亿。他自己的说法是以前从没见过，这就是泡沫。
- 很多企业想 own 自己的 intelligence：基础模型厂商只给你 Pareto frontier 上有限的几个点（大中小几档），但企业自己的 workload 需要的往往不是这几个点，所以要自己训、自己部署、专门为自己的 workload 优化。这也是为什么他说专用架构的 lab 在跑赢通用的 lab。
- 应用层公司每 4 个月就要 reinvent 一次自己，因为底层模型一直在变，地板一直在抬。
- 组织从 hierarchy、command and control 往 network of agents 走：AI 负责信息流转，人相对自主地工作。他说没有 Jack Dorsey 几个月前那篇帖子说得那么夸张，但方向就是这个。
- labs 内部：模型去年大部分时间已经在自己造自己，最近才意识到 alignment 没被当回事；都在押 custom silicon、新架构和 continual learning，算力全员紧缺；为了抢 API token，一边降价，一边派 consultant 去给 Fortune 500 定制吃 token 的工具
- 模型能力和实际落地之间有很大的 diffusion gap，这就是应用层的机会。
- hyperscaler 今年开始借钱做 capex，不再靠 free cash flow。
而且我特别喜欢他这种说话方式，他每句话就是只说一遍，说过去就过去，所以信息密度很高。推荐大家可以直接去看原视频，
总共视频时长只有 15 分钟：https://t.co/HMumZu4bgD > 引用 @gradypb: The @BostonCollege Investment Committee (an LP and my beloved alma mater) asked for a few thoughts on what's happening in AI. I recorded a test run yesterday morning and then shared it with my partners, who encouraged me to share it more broadly... so here you go!
This is not a sales pitch, it's just a reflection on what we're seeing. And it wasn't intended to be shared, so please pardon the rough edges.
https://t.co/AuZqHMSSxD https://x.com/0xTykoo/status/2103863836000522255
## @TheBigBerbowski (TheBigBerbowski) · 09-26 12:58 · ♥38 ↻2 💬3 Micron Technology $MU: A critical layer powering AI https://x.com/TheBigBerbowski/status/2103831599510884844
## @LoganJastremski (Logan Jastremski) · 09-25 19:24 · ♥31 ↻4 💬5 Dropping a podcast with @pequityresearch
Mr. P has been doing some great work breaking down the AI buildout and following where hyperscaler capex actually goes. In this episode I wanted to walk through the full stack with logic, memory, power, and networking.
We get into why memory could become the largest line item in the AI bill, what old GPU rental prices tell us about compute demand, and where value is going to accrue as the physical constraints get harder to solve.
At the center of the conversation is his view that AI spending cannot grow forever. Long-term contracts might change the shape of the next memory downturn, but they don’t eliminate the cycle. And signing a 10-year contract does not mean anyone can actually see 10 years of demand.
We also get into why he’s excited about optics, where NAND and HBF fit as agents use more memory, and his views on Chinese open source and the future of US model development.
We discuss: - Why he thinks 10-year demand visibility is bullshit - Memory’s growing share of hyperscaler capex, and why estimates vary so much - Why older GPUs are still renting and what that says about compute demand - How long-term agreements, pricing floors, and prepayments actually work - Power as a bottleneck, and why identifying a constraint isn’t the same as finding an investment - Copper vs optics, and where networking value accrues - Agents, NAND, and where HBF fits in the memory hierarchy - CXMT, Chinese open source, and the risks of slowing frontier model development
Timestamps: 0:00 – Why AI Spending Can’t Grow Forever 1:13 – P Equity Research’s Background 5:48 – Where Hyperscaler Capex Actually Goes 8:00 – Is Compute Still Tight? 12:08 – Memory’s Share of the AI Bill 17:20 – Why the Memory Cycle Isn’t Dead 18:37 – Inside a Long-Term Agreement 28:30 – What Happens If Customers Cancel? 32:17 – The Rising Cost of the AI Buildout 33:43 – Copper vs Optics 38:49 – Power and Gas Turbines 39:29 – US Models and Chinese Open Source 42:36 – Agents and NAND 47:00 – Where HBF Fits 52:23 – What P Is Most Excited About 56:27 – CXMT and China’s Memory Industry 1:04:00 – Closing Thoughts
Enjoy! https://x.com/LoganJastremski/status/2103566276723613791